Agentic Signal — DeepSeek R1

⚠️ Experimental Model

This model is currently experimental and primarily intended for testing and research inside the Agentic Signal ecosystem.

While it can successfully generate workflows and structured outputs in many cases, it does not yet reliably produce valid workflows every time and may occasionally generate malformed or incomplete node configurations.

Expect inconsistencies, hallucinated structures, and occasional formatting issues. Further fine-tuning and dataset improvements are planned.

Fine-tuned DeepSeek R1 model optimized for structured tool calling, workflow automation, and local AI agents inside the Agentic Signal ecosystem.

Built for reliable JSON output, agent workflows, node orchestration, and local-first AI execution using Ollama.

✨ Features

• 🧠 Optimized for Agentic Workflows
• 🔧 Improved Tool Calling Reliability
• 📦 Better Structured JSON Responses
• 🖥️ Fully Local Execution
• 🔒 Privacy-Friendly Inference
• ⚡ Ollama Ready
• 🔗 Designed for Agentic Signal Integration

🚀 Quick Start

Pull the model:

ollama pull code-forge-temple/agentic-signal-deepseek-r1

Run the model:

ollama run code-forge-temple/agentic-signal-deepseek-r1

🧩 Recommended Usage

This model is specifically tuned for:

• AI workflow automation
• tool/function calling
• structured outputs
• JSON generation
• multi-step reasoning
• local AI assistants
• node-based orchestration systems

Works with:

• Agentic Signal
• Ollama

🔗 Agentic Signal

This model is part of the larger Agentic Signal ecosystem:

Visual AI Workflow Automation Platform with Local Agent Intelligence

Build visual workflows with:

• local LLMs
• AI agents
• tool calling
• structured data processing
• workflow orchestration

without requiring cloud AI providers.

📚 Links

• 🌐 Website & Docs: https://agentic-signal.com
• 💻 GitHub Repository: https://github.com/code-forge-temple/agentic-signal
• 💬 Discord Community: https://discord.gg/FpT2VFdFYu
• 💬 Reddit: https://www.reddit.com/r/AgenticSignal/

❤️ Support the Project

If you enjoy the project, please consider:

• ⭐ starring the GitHub repository
• 🧪 testing the model
• 🐛 reporting issues
• 💡 suggesting improvements
• 🤝 contributing to development

Base Model & License

This model is a fine-tuned derivative of DeepSeek-R1 by DeepSeek AI.

Original license: - MIT License

This repository includes modifications and fine-tuning performed by Code Forge Temple.

Built with ❤️ by Code Forge Temple